Updated: Mar 5, 2020
This episode features Irina Farooq, Chief Product Officer of Kinetica, a Silicon Valley-based company that develops a distributed, in-memory database management system using graphics processing units. Irina shared in-depth insights on how to deploy AI at scale as well as inspiring career advice for product managers.
Irina Farooq is Chief Product Officer for Kinetica, an active analytics leader in the Extreme Data Economy. Irina has over a decade of product management experience across a variety of sectors, including enterprise software, networking, hardware, IoT, SaaS, and Cloud. She previously held leadership roles in Riverbed Technology and Grid Net. Irina has bachelor’s degrees in Mathematics and Computer Science from MIT, as well as an M.B.A. from Stanford University.
[Interview Highlight 1] Machine Learning: From Lab to Production
[Interview Highlight 2] Career advice for entry-level and mid-level product managers
Robin.ly is a content platform dedicated to helping engineers and researchers develop leadership, entrepreneurship, and AI insights to scale their impacts in the new tech era.
Subscribe to our newsletter to stay updated for more inspiring talks:
1. The role of Chief Product Officer
Margaret Laffan: Irina, we are delighted to have you with us today.
Thank you for having me.
Margaret Laffan: We're going to talk a bit about your career, and about how things are over at Kinetica, and how you're doing there, your role as a Chief Product Officer. And I'm really looking forward to having this discussion today because I know you've had a very exciting career to date. Tell us a bit more around that: You started your career as an engineer after graduating from MIT, and then quickly transitioned to product management. What triggered your transition?
It was definitely, absolutely by chance and stroke of luck. I actually transitioned into product management just three months out of college. So I joined Oracle and I thought I was God's gift to mankind. And everybody was waiting for me to solve the hardest problems. And then I realized that I was one of 500 engineers on the project. And I kept asking people, why are we doing this? Why is this so important? And I realized that I was going to have to work for a long time to be able to answer those questions. And so I got some good advice from somebody who said, actually I think you would be better in this other role. And I'm like, what's that? It’s pretty risky because as an engineer, you worry about losing your technical skills. But overall, it's been absolutely the best transition for me.
Margaret Laffan: We're going to jump into that and dive a bit deeper in terms of what it means to be a Chief Product Officer. Because you hold multiple different hats. You're responsible for all product-related matters. This obviously includes your product vision, especially at Kinetica - product innovation, product design; you look at product management, project management; you're talking also around product marketing. Can you help us understand what your day to day looks like?
At Kinetica, I'm responsible for product management, as you said, product marketing, and program management to help build the machinery of delivery of software. And my day could be ranging from working on the launch of a new product to working with one of our largest customers to either figure out a new use case, or sometimes figure out an issue, or helping guide my team to work towards the next phase of innovation. So it varies from both driving the organization along the right process and technology vision that we'd like to drive, as well as engaging with our customers, partners, analysts to help make that externally a reality.
Margaret Laffan: I would imagine that every day is pretty hectic. And every day, there's something different happening as well. You compartmentalize times during the week where you spend time learning or you spend time coaching or mentoring with your team. Can you tell us a bit more around how you manage your time?
It's interesting. One thing with a job like this is very easy to get sucked into the vortex of urgent things. And the worry is that you don't get to focus on the important things, if you just focus on the urgent.
So what I do, actually, every Sunday night, I set goals for myself and the organization for the next week. And I think that helps me by the time I come into the office, and we have people in all time zones around the world, my mailbox is full. I at least know what's important to me, and I can prioritize against what I'm driving all the urgent fire drills that are coming in. And then the other thing with my team is, I kick off a team meeting with my team first thing on Monday, to make sure we're all aligned. And then towards the end of the week, I have one-on-ones with each of my direct reports to make sure we close the loop on those things. So on Monday, I know the teams are aligned and they're driving specific things that we've all agreed on. Then on Friday, we take a step back and see what worked, what didn't. And it also gives them an opportunity to voice things that may not be appropriate in a public forum.
2. Team Collaboration
Margaret Laffan: When we think about the Chief Product Officer role, this is newer for tech companies in Valley. From a leadership and CTO perspective, how do you and your colleagues work together in terms of understanding? What each of you are responsible for, and then how you would collaborate together?
I have a little bit of an advantage at Kinetica. Our CTO and founder, who is amazing, he picked me to be his peer, so we came in with an understanding of that partnership. But in general, I work very hard that I have to, as a leader, work with my immediate peers, to make sure we're aligned and there is no politics between us. So our teams can work as a single team driving the same objectives.
So with our CTO, we have one product tech, one product vision, we present jointly to the board, because the moment product and engineering starts giving different narratives, you know you're not in a good place. Similar to our marketing team, while I own product marketing, making sure that I'm aligned with the marketing team, and the CMO and all the things we're driving, because I can't be everywhere. The teams need to collaborate freely and feel safe, and feel like they can take decisions independently.
Margaret Laffan: Right. And for a company that's scaling so fast, I can imagine having to do that extra work around collaboration, ensuring you're all aligned, it takes good synergy, it takes good vision, settings and so forth.
Yeah, I think it's reinforcing the principles and part of leadership is like, you have a very clear picture about where you want to go and you articulate it, but getting there, that’s really hard. No matter how great the place you're at it, you have to go through a lot of slog to get it, because as you're driving organizational change, evolving people's roles, you're introducing new people into the organization, as situations arrived, having clear vision as a leader - this is this person's job, this is your role, this is how you communicate - and reinforcing what your stated goal has been. In my mind, I do a lot of that. What are our principles? How do we act and reinforce the same message over and over again, so the team feels both safe and empowered?
3. Advice for Entry-level PM
Margaret Laffan: For those who are listening and are interested in becoming better product managers, and are taking on these roles, what advice would you have to share with entry-level PMs?
It depends on where you're coming from. But a lot of people come into product from an engineering background. And it's not an easy transition, I won't lie. And part of it is not because you don't have the jobs, of course, you do, most people do. It’s more because your context is so widely different. When I transition people into the product, I tell them, you no longer have the luxury to live in your own head. You have to live in other people's heads. You have to live in the heads of your customers, and not just customers you sell the product to, but it’s engineers, it’s sales, it’s marketing. It's you’re now all of a sudden, at the nexus of a lot of people who need you have to understand and help drive that.
And so it's not an easy transition, but it's a very valuable one. Because even if you progress in any other function as you become more senior, I believe these are the skills you will have to learn. In product, you get to learn influence without authority, which I personally believe is a better model than exercising authority every time to get people to do the right thing or the thing you are passionate about,
Margaret Laffan: We talked about the entry-level, but what about mid-level PMs who are looking for that next experience jump or next promotion? What type of recommendations would you have for them?
Irina Farooq: It's interesting, I would say it applies in every function. I think there's different type of promotions. There is gradual like you're a product manager, you become a senior product manager; you’re a director, then you become a senior director. And then there’s a step function promotions, you go from individual contributor to a director; you go from a director to VP, VP to C-level. Those are step functions where your context becomes 10x the context you had in the previous role. So it's not sufficient anymore to just do your job really really well and guarantee to be promoted to the next level.
My recommendation to that is - I've never asked for promotion in my career. What I've asked for were experiences. Because as I look at people for the new organization, I think about what do they know about how to do that I don't? And how can I learn those things? And also, how can I help the organization with things that nobody's handling? I don't recommend go take a project somebody else is doing and ask for it. But there's usually a lot of things you can do to help the organization that nobody has the capacity to do. There's nobody in that role. And asking for experience and showing leadership on that is usually the best way to get the skills. And then titles equalize, once you've shown that type of leadership potential and driven some initiative that is beyond your current level.
Margaret Laffan: It’s a good question to ask yourself around the value you can add. Where can you help the organization move to that next level? And then what sort of problems do you see that your peers or your leaders struggling with? And then how can you help them solve something or do it better? I really like that approach on the experience side.
I used it when people in my team came to me and said, I want to get promoted. It's a conversation, but let's go back to your goals. What are your goals? What are you trying to accomplish? Some people's goals is simply to get promoted. And as a leader, I have to be transparent. This is the Delta. These are the experiences that you need to get to that next level. And I can help surface those experiences, and I can say, I'll build the roadmap for you and how to give you exposure and coach you through it. So then, when you're at that level, you don't sink or swim, you're prepared to actually take on that additional responsibility.
4. Why Kinetica?
Margaret Laffan: Why did you choose Kinetica after your four years at Riverbed Technology?
It's very interesting because Kinetica is probably the first company where I came with an understanding of customers’ needs a priori because I would have been the customer. Because as a product person in past years building Smart Grid solutions, software-defined networking, I tried to build this distributed system that relied on analytics to make decisions in real-time.
And what I realized that every time I come to my engineers who were incredibly bright, the best engineers, and they were like, to add this feature takes two months. But I wouldn't understand because it felt simple. And as I started digging, I actually understood that it was the data layer that was the challenge. A lot of the things that the notion of analytics is changing from static reports to this running artificial intelligence in near real-time to be able to make decisions. And the data layer was the bottleneck. So Kinetica was uniquely solving that problem. And I was like, if I had this, I would have built completely different products before, so I want to build this and help everybody.
Margaret Laffan: How would Kinetica differentiate itself from its competitors from a technology point of view?
There's a lot that we've done. When you think about this new type of applications, let's say, you call Uber. I live in San Francisco, I use Uber to get to work. You have lots of streaming data that comes in from all the drivers, weather and traffic, etc. And when I call Uber, it gives me a prediction of how long it will take me to get to work, how much it's going to cost me, when the drivers are going to be there. That's taking all that real-time data, cross-correlating it against historical patterns, and running predictive models to give you an output in real-time. Everybody in their specific industries wants to build their version of Uber that will advance that customer experience and deliver those outcomes.
At Kinetica, we built a platform that helps you do that, whether you're a developer, a data scientist, or an analyst. We've done a lot of things to simplify that architecture to give you the tools to build those applications.
5. Machine Learning Application: Strategies and Cost
Margaret Laffan: When you think about building those applications and obviously leveraging machine learning, there’s a lot of conversation around how machine learning is still research-based, is still in the lab, hasn't gone into full production, hasn't some cases and then others, it's still on the way. What’s your approach, your execution strategy or tactics that you use to bring machine learning out of the lab into production?
It's interesting. We see to your point, we work with a lot of customers and a lot of things we built into the product to help ease the pain of taking AI out of the lab and into production. But it's not just a technology problem, it’s also a "people in a process" problem. Because if you have data scientists in an isolated team, and they're looking at data and building models, but they don't have business-level sponsorship, they're not aligned with the business or product managers within the business, and what they're driving, then everybody gets frustrated.
So making sure that you have the right people, data scientists, data engineers, product managers, engineers working together, integrated into the same software development teams is really what we've seen is a big accelerator to getting AI out of the lab into production.
On our end, what we're focusing on the technology aspect of that is how do you do it if you have the models that align with your vision, how do you deploy AI at scale in a way that's responsible and protect you, that you're able to govern and protects you against bias, and other things we see in the marketplace.
Margaret Laffan: Let’s talk a bit more around the cost of all of this, because we keep coming back to that. You've got your team in place, there's still some research you're bringing into production at scale and deployment at scale. What does it take to be production-ready in this current market, can you give us a sense of how long that research cycle should go? How would you accelerate that process?
It depends, I think different companies are in different level of maturity on that. But I think today, in general, to get to “Hello World”, to get to your first application is really time-consuming. It's time you have to put a lot of different technologies together in order to be able to get to that “Hello World” point. But I think that's the obvious initial cost. I think the problem is, there are a lot of hidden costs that companies don't get now. Because we talked to a lot of Chief Data Officers and they say, I don't really know which models I have run anywhere. Now imagine we have the headlines about the Apple Card and the bias there. If you don't know what models you have, how can you monitor them for bias? How do you monitor them for accuracy? How do you monitor their performance over time? I think 2020 will become the year that we'll get to that maturity around how they deploy this application.
6. 2020 Goal and Future Direction
Margaret Laffan: For Kinetica in 2020, what have you focused on?
At Kinetica, we're extremely fortunate. We focus on the Fortune 2000 and federal customers. It's almost like Kinetica makes us relevant at parties, because we get to see all the transformations that some of the largest enterprises in the world are going through. I do think that 2020 for many of them will become the year to really start going in that journey of bringing machine learning out of the lab and into production in a much more thoughtful way, versus experimental projects. People are now getting to the point where they're putting guidelines and best practices, taking into account the different stakeholders.
Margaret Laffan: It's maturing. So you would anticipate to see a lot of that in 2020.
Margaret Laffan: From a leadership perspective and a new leader in high tech, what's been changing regarding the talent market in the tech industry? What sort of shifts have you seen in the last few years, especially when it relates to machine learning, to talent that's coming out of the different universities or the experience that people are getting and scaling in growth companies? And what are your observations around that?
It's interesting, because the previous generation of innovation has opened about software and building software. And the new generation is about AI, which is really data. So I think no matter which industry you’re in, you may not be directly the AI guru, but you need to have a fundamental understanding of the data, both data technologies, but also what it means to use data for predictive analysis. I think the bar of understanding of data and machine learning technologies is getting higher and higher.
I work a lot to make sure I get candidates from diverse backgrounds with different levels of seniority. But you see a lot of interesting people coming out of universities, they barely have any experience, but they have such relevant knowledge and specific domains, that sometimes they get an edge over more experienced candidates, because they have an understanding of these problems. It used to be like in 1990s and 2000s, the enterprises didn't care about user experience that you were used to. Usability was just like reflecting how the data is stored. That has changed significantly. Consumer internet has changed that, because it is bled into the enterprise. And it changed the expectation. And I think AI now puts the next level we’re going to focus on. And in a lot of ways, it's a lot about the new types of experiences that we could provide to customers. So I think you're right, it's around that kind of mirroring the previous UX focus that we brought in, and then data becomes the next element.
Margaret Laffan: Which is why I also love very diverse workplaces and diverse teams, because you need a lot of different thought processes and brain trust into that to create the next generation of experience that you're looking to do.
That's interesting because sometimes people on my team come, well, how will I work with this person? We have such different experiences Why do you think you assigned us to work together? That's the beauty, right? Yes, there's a little bit of language building upfront, but that cost is worth it to get to the right outcome.
7. Leadership Style and Values
Margaret Laffan: I'm very curious around how you would describe your leadership style. We know there's affiliative, democratic and visionary styles. What are your leadership style and your values?
It's interesting. I think the number one value for me is respect. I've been willing to lose my top performers over cases where they haven't shown respect and drawn clear lines on that. The other one is open-mindedness. When you are solving problems, you can't be closed to solutions that you haven't thought about. And so having an open communication and collaborative culture has been very important to me. I put culture fit first, over anything else, over skill, overpotential, over anything.
Margaret Laffan: Irina, what’s the next chapter look like for you?
To grow Kinetica to be a household name in data analytics. We're thankful to be in an amazing trajectory and I hope that continues. I get the unique opportunity here to go to the product organization and the product vision from early on. But the next after this, I'm hoping that I'll be running a company in the technology enterprise space.
Margaret Laffan: Awesome. Taking that next level of experience right after everything you've been through to take it to this stage.
It would be the next level of experience, and having the right mentorship. For example, now, even our CEO exposes me to things that are above my job, knowing that that's eventually my aspiration. And so that's going back to, I'm not going to get promoted to Kinetica, I love my job, I love my team. This is the job here. But knowing that just set myself up for that next chapter, whenever it comes, I need to get exposure to all these other things that are between me and the CEO today.
Margaret Laffan: I love that time and place experience, and you look for it and you seek it and you're on that journey. Irina, we cannot wait to hear and see where you go next and wish you very good luck in your next adventure.
Thank you very much.